Within artificial intelligence, the sub-field of multi-agent systems studies the foundations of agent interactions and strategic behavior. Two-sided matching is one of the most fundamental problems in this field with applications in matching residents to hospitals, kidney donors to receivers and students to high schools. The earliest algorithm that solved this problem is the Gale-Shapley algorithm which guarantees a stable matching based on the preferences of both sides but has a drawback of favoring one side over the other, that is, proposers
always get their most optimal stable partner.
We consider the design and analysis of gender-neutral stable matching algorithms where the proposing side from both sides is randomly chosen thereby giving an equal probability for both sides to get their most optimal stable partner (ex-ante). Later, we focus on investigating if an agent can exhibit strategic behavior i.e., whether it is possible for an agent to manipulate so that he/she improve the partner obtained when on the proposed side while retaining the partner obtained when on the proposing side.
The results obtained showed that for some manipulation algorithms, agents can still manipulate the outcome even when the decision of which side is proposing is unknown. Also, empirical evaluations were performed to understand and solidify the results.
Library of Congress Subject Headings
Multiagent systems; Matching theory; Computer algorithms
Computer Science (MS)
Department, Program, or Center
Computer Science (GCCIS)
Rudraraju, Sanjay Varma, "Strategic Behavior and Manipulation in Gender-Neutral Matching Algorithms" (2018). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus